The XO Laptop: It’s the Software, Stupid

Wade Roush2/1/08Follow @wroush

Share and Comment

24 Comments Order a Reprint E-mail this Story

• Share on Twitter
• Share on Facebook


• Share on Linkedin
• Share on Google
• Submit to Hacker News

On YouTube, there is an 11-minute video of the veterinarian-assisted birth of a calf on a farm in Villa Cardal, Uruguay, a small town in a dairy-rich region four hours north of the capital, Montevideo. It’s an amazing thing to watch—at least, to a city slicker like me who doesn’t get to witness the miracle of birth every day. But what makes this particular video remarkable is that it was shot by a fourth-year student at Villa Cardal’s Public School 24, using the built-in camera and recording software on the student’s XO Laptop, within weeks of the machine’s arrival at the school last year.

Uruguay was the first country to purchase a large number of XO laptops, ordering 100,000 of the small green machines from Cambridge, MA-based One Laptop Per Child Foundation last October. If you dropped a couple thousand bucks on your last laptop, you may be alarmed by the idea of a student taking her brand-new XO into a muddy cow pen and getting up close and personal with a caul-enmeshed calf still shining with amniotic fluid. But to the folks at OLPC, who designed the $175 XO to be rugged and portable, yet powerful, finding the YouTube video was a triumphant moment. This bit of barnyard reality spoke volumes about an often-overlooked aspect of the project—namely, the software, which is designed to overturn old notions of classroom learning and give kids the ability to collaborate and express themselves in many media.

“I was in Brazil, at home, and it’s around 1 o’clock in the morning on a Saturday, and being an idiot, I’m on e-mail,” says David Cavallo, OLPC’s chief learning architect and the former Latin American coordinator for the project. “And I get a note from a regional coordinator—who reports directly to the president of Uruguay—who sends me this video from Villa Cardale, where there are 150 kids and every kid got a laptop. Nobody taught them how to do this, but they’re already making their own stuff and posting it to YouTube! You can see this fluency developing, a sense of what it means to express something in video. It’s really quite articulate.”

OLPC and the XO Laptop have received mountains of media attention over the past year, most of it focusing on the laptop itself, the foundation’s difficulties getting the device into mass production and lining up solid orders, and OLPC founder Nicholas Negroponte’s public clash with Intel (a saga we hope we have helped put to rest with our January 25 analysis and our January 28 interview with Negroponte). In one recent post, CNET blogger Tom Krazit complains about delays in the delivery of XOs to donors who participated in OLPC’s recent “Give One, Get One” program (See “Give one laptop, get one sooner or later“); while Krazit’s blog is usually valuable for its level-headed coverage of Apple, his XO post typifies the punditocracy’s skeptical, often mystifyingly angry and dismissive stance toward OLPC and the XO.

What almost all of the coverage of OLPC has omitted—and what came out over and over in my recent interviews with Cavallo and OLPC software president Walter Bender—is that the XO does not really matter as a piece of hardware. It matters a great deal as a vehicle for a very specific vision about learning, education, and the possibilities for engagement through digital technology. And that vision is much more evident in the software that comes pre-loaded on the XO than in the physical design of its hardware.

The hardware, it must be said, seems almost expressly designed to baffle and frustrate adults; in the name of durability and portability, the designers have thrown out most of the comfortable conventions of laptop design, evoking predictable reactions from the computer industry’s usual media gatekeepers. With icy British sarcasm, one Economist reviewer (in a January 4 article entitled “One clunky laptop per child“) derided the machine as having buttons too small for adult hands, a frustratingly slow processor, a cumbersome and buggy operating system, a “screwy” track pad, and a keypad that (horrifyingly) “lack[s] the normal press response that allows smooth typing.”

Bender, who runs the software side of the OLPC Foundation and designed the XO’s top-level interface, called Sugar, admits that the laptop wasn’t designed for “fat-fingered adults.” (It was designed for elementary school students in the 6-to-12 age bracket; see Bender’s point-by-point response to the Economist review here.) “They’re right—it is a clunky laptop, depending on what your metric is,” Bender told me when I visited OLPC’s Kendall Square offices two weeks ago. “But my response is that they’re using the wrong metric. It doesn’t matter whether it’s got a faster or slower boot time than this or that machine. Are the kids learning? That’s the only metric that matters.”

And by that metric, it’s probably too early to render a verdict, since the XO went into mass production less than three months ago and only about a quarter-million of the devices have shipped so far. But as the Uruguyuan birthing video demonstrates, students are already putting the XO’s media-creation tools to creative use, bearing out at least one of the OLPC’s five core principles—that of connection, the idea that children should be able to use the XO to express themselves in communication with one another and with the larger world via the Internet. (The other four principles are that every child should have his or her own laptop, that students should be able to own the laptops and take them home after class, that the program should focus on elementary-school students, and that all of the software and systems on the machine should be free and open-source.)

Bender doesn’t expect the XO to give rise to legions of young YouTube-posting videographers; for one thing, broadband connectivity is in scarce supply in the rural, developing regions where OLPC has focused its distribution efforts. But students can still use the laptop to connect with one another, via the machines’ built-in Wi-Fi mesh networking system. In fact, when you turn on the XO, the first thing you see isn’t a list of applications—it’s a menu leading to a map of the other people using XO’s in the neighborhood, wirelessly speaking.

“We knew that Internet bandwidth was going to be available but expensive, so that having all these kids go off to MySpace [to do their networking] wasn’t necessarily viable,” says Bender. “We needed to capture that collaboration and bring it closer. So we thought, why not just make collaboration a seminal part of the operating system? Instead of having collaboration be something that happens out there—something where you graft network awareness onto other activities—we said, let’s have it be something that’s always part of what you do. So we put the notion of the presence of other people in the local area network directly into the interface.”

Collaboration is also a central theme in many of the applications Bender’s team chose to include in Sugar. The best-developed example so far, Bender says, is the XO’s music software suite. Working with Jean Piché, a composer and music producer who teaches electroacoustic composition at the University of Montreal, Bender’s team first assembled a simple application called TamTam that allows students to create music by choosing from hundreds of synthesized sounds, then pounding the keys. “From there you go to TamTam Jam,” says Bender. “It’s network-savvy, and it will synchronize with other laptops, so you can turn on rhythms, access multiple instruments at once, and build rich musical structures.” Then there’s TamTam Edit, a sequencing and composition tool that lets students transplant tracks from one composition into another over the network, and TamTam synthLab, which lets them invent new sounds—for example, by applying random number generators to a library of recorded waveforms.

“But wait, there’s more,” Bender says. From synthLab, students can drill down into Csound, a musical scripting language invented by Barry Vercoe at the MIT Media Lab that’s so powerful that film-score composers use it to create special musical effects. “So you go from a tool that a two-year-old can immediately start using to the same tools they use in Hollywood,” says Bender. “Not every kid is going to be a Csound hacker, but there is this opportunity for growth, exploration, invention, expression, sharing, and critique that is part of everything we do.”

Hang out at the foundation very long, in fact, and you’ll realize that Bender, Cavallo, Negroponte, and their colleagues are utterly sincere when they say OLPC is a learning project, not a laptop project. The XO is simply an embodiment—the best one current technology will allow, given the compromises necessary to make the device cheap, durable, and power-miserly—of constructivist learning principles that philosopher-psychologist Jean Piaget and computer-science pioneers like Seymour Papert and Alan Kay have espoused for decades, and that have long had a home at the MIT Media Lab, which, of course, Negroponte himself founded in 1985.

“There is a long history of technology and learning,” says Bender. “We and others have been working on this problem for almost 50 years. So we have a Logo environment on the XO, called TurtleArt. [Logo is the graphical programming language co-invented by Papert.] We have eToys, a fabulously rich program that is the embodiment of Alan Kay’s life work.” In fact, almost every piece of software on the XO is designed to advance the constructivist belief that learning occurs most efficiently when it’s active, social, and exploratory, with constant feedback between instructors and learners and between learners themselves.

Sugar does include some unique quirks—”things we’ve done that we’ve done because we thought they were necessary, and because we had the opportunity to correct things that were just wrong in other systems,” in Bender’s words. For instance, Bender has a grudge against double-clicking and overlapping windows, two user-interface conventions that he says are “fundamentally bad ideas” and are notably absent from the XO. But “we are not reinventing the wheel if we can help it,” he says. The operating system on the XO is Fedora, a free version of Red Hat Linux; its Web browser is a version of Mozilla’s Firefox; the word processor is based on Abiword, a popular open-source program; its streaming media player is a free version of RealNetworks’ Helix system.

Bender says his goal for 2008—aside from simply getting more laptops into the hands of students, of course—is to make the Sugar platform more stable and get the remaining bugs out. “We also want to do a better job of supporting the community, so that more flowers can bloom,” he says. That’s the main reason that virtually everything on the laptop, right down to the hardware drivers, is open-source—so that it can be shared and so that, ultimately, responsibility for maintaining the platform can be transferred from the foundation itself to the community of educators, students, and developers using the XO. “In open-source you strive to push everything upstream, because as soon as it’s upstream, it’s not your problem anymore, it’s the community’s problem,” says Bender. “That’s a great place to be. And we are trying to push as much upstream as possible, because we won’t be successful otherwise.”

Especially not once shipments scale up to the millions, and students are using the laptop in dozens of countries, including the United States. Ultimately, the XO Laptop will succeed only in proportion to the creativity that students, teachers, software developers, and education ministries pour into it—and that’s largely out of OLPC’s hands. But the flowers are already starting to bloom, on YouTube and elsewhere. (The blog for Uruguay’s Ceibal Project, whose official goal is to give one laptop to every child in the country, is full of examples like the cow video).

“Now, are we finished?” asks Bender. “No, we’ve got a long way to go. But it’s beginning to happen. The laptops are getting into the hands of kids and into the hands of software developers, and they are starting to use them and discover things they like and things they don’t like, and they are fixing the bugs, and they are supporting each other. It has exceeded my expectations, the level to which the community has engaged in the process”–and assisted in the XO’s own birth.

Wade Roush is Xconomy's chief correspondent and editor of Xconomy San Francisco. You can subscribe to his Google Group or e-mail him at wroush@xconomy.com. Follow @wroush